Basic Data Visualization: topic introduction

In this part of the course, we will cover the following concepts:

• Discuss basics of Univariate, Bivariate & Multivariate plots and their use cases
• Build plots with base r

Chat question

• Exploratory Data Analysis refers to the analysis process which is used to discover patterns, spot anomalies, test hypotheses, and check assumptions
• During EDA, you will likely generate visual representations of data to see trends, outliers, and patterns
• In your own work, how do you approach exploratory data analysis?
• Share your responses in the chat

Module completion checklist

Univariate plots

• Univariate plots are visualizations that represent a single variable
  • These include boxplot, histogram, line plot, density curve, dot plot, QQ plot, bar plot
• It can be helpful for profiling a dataset in the initial stages of EDA to learn more about individual variables
• Combining univariate plots to compare the distributions of different variables can also be done

Directory settings

• In order to maximize the efficiency of your workflow, we use the box package and encode your directory structure into variables
• Let the main_dir be the variable corresponding to your materials folder

# Set `main_dir` to the location of your materials folder.

path = box::file()
main_dir = dirname(dirname(path))

Directory settings (cont’d)

# Make `data_dir` from the `main_dir` and
# remainder of the path to data directory.
data_dir = paste0(main_dir, "/data")
# Make `plots_dir` from the `main_dir` and
# remainder of the path to plots directory.
plot_dir = paste0(main_dir, "/plots")

Case study: stroke survey

Stroke Dataset: attribute information

• id: unique identifier
• gender: “Male”, “Female” or “Other”
• age: age of the patient
• hypertension: 0 if the patient doesn’t have hypertension, 1 if the patient has hypertension
• heart_disease: 0 if the patient doesn’t have any heart diseases, 1 if the patient has a heart disease
• ever_married: “No” or “Yes”
• work_type: “children”, “Govt_job”, “Never_worked”, “Private” or “Self-employed”
• Residence_type: “Rural” or “Urban”
• avg_glucose_level: average glucose level in blood
• bmi: body mass index
• smoking_status: “formerly smoked”, “never smoked”, “smokes” or “Unknown”*
• stroke: 1 if the patient had a stroke or 0 if not

Load the dataset

• Let’s load the dataset from our data_dir into R’s environment

# Read CSV file called "healthcare-dataset-stroke-data.csv"
health_data = read.csv(file = file.path(data_dir,"healthcare-dataset-stroke-data.csv"), #<- provide file path
               header = TRUE,            #<- if file has header set to TRUE
               stringsAsFactors = FALSE) #<- read strings as characters, not as factors

View data types

• Let’s examine the data types of the columns in the dataset and handle the missing data, if there are any

str(health_data)

'data.frame':   5110 obs. of  12 variables:
 $ id               : int  9046 51676 31112 60182 1665 56669 53882 10434 27419 60491 ...
 $ gender           : chr  "Male" "Female" "Male" "Female" ...
 $ age              : num  67 61 80 49 79 81 74 69 59 78 ...
 $ hypertension     : int  0 0 0 0 1 0 1 0 0 0 ...
 $ heart_disease    : int  1 0 1 0 0 0 1 0 0 0 ...
 $ ever_married     : chr  "Yes" "Yes" "Yes" "Yes" ...
 $ work_type        : chr  "Private" "Self-employed" "Private" "Private" ...
 $ Residence_type   : chr  "Urban" "Rural" "Rural" "Urban" ...
 $ avg_glucose_level: num  229 202 106 171 174 ...
 $ bmi              : num  36.6 NA 32.5 34.4 24 29 27.4 22.8 NA 24.2 ...
 $ smoking_status   : chr  "formerly smoked" "never smoked" "never smoked" "smokes" ...
 $ stroke           : int  1 1 1 1 1 1 1 1 1 1 ...

Impute missing data

• We will now impute missing values in the bmi column with the mean

# Convert BMI to numeric
health_data$bmi <- as.numeric(health_data$bmi)
# Replace N/A's in BMI column with mean
health_data$bmi[is.na(health_data$bmi)] <- mean(health_data$bmi,na.rm=TRUE)

str(health_data)

'data.frame':   5110 obs. of  12 variables:
 $ id               : int  9046 51676 31112 60182 1665 56669 53882 10434 27419 60491 ...
 $ gender           : chr  "Male" "Female" "Male" "Female" ...
 $ age              : num  67 61 80 49 79 81 74 69 59 78 ...
 $ hypertension     : int  0 0 0 0 1 0 1 0 0 0 ...
 $ heart_disease    : int  1 0 1 0 0 0 1 0 0 0 ...
 $ ever_married     : chr  "Yes" "Yes" "Yes" "Yes" ...
 $ work_type        : chr  "Private" "Self-employed" "Private" "Private" ...
 $ Residence_type   : chr  "Urban" "Rural" "Rural" "Urban" ...
 $ avg_glucose_level: num  229 202 106 171 174 ...
 $ bmi              : num  36.6 28.9 32.5 34.4 24 ...
 $ smoking_status   : chr  "formerly smoked" "never smoked" "never smoked" "smokes" ...
 $ stroke           : int  1 1 1 1 1 1 1 1 1 1 ...

Univariate plots: line plot

# Line plot.
plot(col_vector,type,col,xlab,ylab)

Univariate plots: line plot (cont’d)

• Let’s create a simple line plot for the bmi column to observe the trend in it

# Line plot.
plot(health_data$bmi,
     type = "o",
     col = "orange",
     main = "bmi")

Univariate plots: boxplot

# bmi summary.
summary(health_data$bmi)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  10.30   23.80   28.40   28.89   32.80   97.60 

Univariate plots: boxplot

# Univariate plot: box-and-whisker plot.
boxplot(health_data$bmi)

Univariate plots: boxplot (cont’d)

# Customize your boxplot.
boxplot(health_data$bmi,
        col = "orange") #<- set color

# Customize your boxplot
boxplot(health_data$bmi,
        col = "orange",
        main = "BMI Boxplot") #<- add title

R base colors

# R has many colors in its native 
# `grDevices` package.
?colors

# There are a total of 657 colors in 
# the `colors()` vector.
str(colors())

 chr [1:657] "white" "aliceblue" "antiquewhite" "antiquewhite1" ...

demo("colors")

Create variable colors

?sample

Create variable colors (cont’d)

# Set random seed to get the same sample every time!
set.seed(1)

# We have as many variables as columns in our data.
# Save number of columns to a variable using `ncol` function.
n_cols = ncol(health_data)

# Pick a sample of colors for each
# variable in our data set
col_sample = sample(colors(),#<- vector of colors
                    n_cols)  #<- n elements to sample 
col_sample

 [1] "dodgerblue1"   "orchid1"       "mediumpurple4" "grey38"       
 [5] "grey9"         "gray34"        "grey46"        "slateblue3"   
 [9] "grey16"        "olivedrab1"    "grey69"        "skyblue2"     

Compare variables: boxplots combined

# Make a box plot of relevant variables and give a vector of colors for each of them.
boxplot(health_data$age, health_data$avg_glucose_level, health_data$bmi, col = col_sample) 

Univariate plots: histogram

# Histogram data without plot.
hist(health_data$avg_glucose_level, plot = FALSE)

$breaks
 [1]  40  60  80 100 120 140 160 180 200 220 240 260 280

$counts
 [1]  220 1312 1599  860  298  153   85  149  229  150   46    9

$density
 [1] 2.152642e-03 1.283757e-02 1.564579e-02 8.414873e-03 2.915851e-03
 [6] 1.497065e-03 8.317025e-04 1.457926e-03 2.240705e-03 1.467710e-03
[11] 4.500978e-04 8.806262e-05

$mids
 [1]  50  70  90 110 130 150 170 190 210 230 250 270

$xname
[1] "health_data$avg_glucose_level"

$equidist
[1] TRUE

attr(,"class")
[1] "histogram"

Univariate plots: histogram (cont’d)

• However, we can also produce a fully visualized histogram

# Univariate plot: histogram.
hist(health_data$avg_glucose_level)

Univariate plots: histogram (cont’d)

# Customize your histogram.
hist(health_data$avg_glucose_level,
     col = col_sample[1]) #<- set color

hist(health_data$avg_glucose_level,
     col = col_sample[1],
     xlab = "avg_glucose_level",   # x-axis label
     main = "Dist. of avg_glucose_level") # title

Compare variables: histograms combined

# Make a combined histogram plot.
par(                                    #<- set plot area parameters with `par`
    mfrow = c(1, 2))                    #<- split area into 1 row 2 columns with `mfrow` 
hist(health_data$age,   #<- add 1st histogram
     col = col_sample[2],
     xlab = "age",   
     main = "Dist. of age")
hist(health_data$avg_glucose_level,   #<- add 2nd histogram
     col = col_sample[3],
     xlab = "avg_glucose_level",   
     main = "Dist. of avg_glucose_level")

Univariate plots: bar chart

• Bar charts are used to represent the data in the form of columns and rectangles
• The syntax of the bar chart is:

# Bar chart.
barplot(input,xlabel,ylabel,main,names.arg,col)

• It includes:

  • input represents the vector or matrix containing the numeric values
  • xlabel and ylabel denote the labels for x-axis and y-axis respectively
  • names.arg represents the vector of names that would appear under the bar
  • col parameter denotes the color

Univariate plots: bar chart (cont’d)

• Let’s create a bar chart with a sample vector of a patient’s pulse rate at different intervals

par(                                    #<- set plot area parameters with `par`
  mfrow = c(1, 1))
# Create the data for the chart
input <- c(80, 95, 90, 85, 75)
time_i <- c("6AM", "9AM", "12PM", "3PM", "6PM")

# Plotting the bar chart
barplot(input, names.arg=time_i, xlab="Time intervals", ylab="pulse rate of sample patient", col="blue")

Module completion checklist

Bivariate plots

• Bivariate plots are visualizations that represent the relationship between two variables
• As was the case when examining single variables, there are several essential characteristics of the relationship between two variables that are of interest:
  • the form of the relationship
  • the strength of the relationship
  • the dependence of the relationship on external circumstances.
• By plotting two variables within the same space, we can add other features, like regression lines, that assist in predicting

Bivariate plots: scatterplot

• One of the main bivariate plots is the scatterplot, representing each observation of two variables as a point in a coordinate plane
• Let’s create a generic scatterplot using one variable as the x and the other as the y

plot(health_data$age, #<- variable for x-axis
     health_data$avg_glucose_level) #<- variable for y-axis

Bivariate plots: scatterplot (cont’d)

• We can build on our base by adding labels and a title

plot(health_data$age, 
     health_data$avg_glucose_level,
     xlab = "age",
     ylab = "avg_glucose_level",
     main = "age vs avg_glucose_level") 

Bivariate plots: scatterplot (cont’d)

• Select a symbol for the observations with pch option from the following options

# Customize your scatterplot.
plot(health_data$age, 
     health_data$avg_glucose_level,
     xlab = "age",
     ylab = "avg_glucose_level",
     main = "age vs avg_glucose_level",
     pch = 8,   #<- type of symbol to use
     cex = 2,   #<- scale of the point
     col = "steelblue") 

Multivariate plots

• Multivariate plots allow you to visualize the data distribution and relationships among three or more variables
• Examples include scatterplot matrix, correlation plot, parallel coordinates plot, various compound plots using multiple juxtaposed variables
• Multivariate plots are most helpful in learning how groups of variables are related, as well as in combination with machine learning methods that require specific complex inputs

Multivariate plots: scatterplot matrix

# Plot scatterplots for many variables.
# Select quantitative variables 
health_quant <- health_data[, c("age", "avg_glucose_level", "bmi")]

pairs(health_quant , #<- select variables
     pch = 19,          #<- set symbol
     col = "steelblue") #<- set color

Installing corrplot package

• A correlation plot is a matrix representing the correlation between variables in each tile rather than the point distribution
• We can use the corrplot package to create these visualizations

# Install package through command line.
#install.packages("corrplot")

# Load the package into the environment.
library(corrplot)

corrplot 0.88 loaded

Multivariate plots: correlation plot

# View package documentation.
library(help = "corrplot")

?corrplot

• The only argument required for corrplot is a correlation matrix of the variables we want the plot to represent
• We can generate this correlation matrix using cor

Multivariate plots: correlation plot (cont’d)

# Create correlation plot.
health_cor = cor(health_quant)
corrplot(health_cor)

Multivariate plots: correlation plot (cont’d)

# The default method is "circle", to
# display the correlation coefficient 
# instead, use method "number".
corrplot(health_cor,method = "number")

# To display correlation coefficient as 
# a portion of a circle proportionate to 
# correlation coefficient use method "pie".
corrplot(health_cor,method = "pie")

Knowledge check

Exercise

Module completion checklist

Basic Data Visualization: Topic Summary

In this part of the course, we have covered the following concepts:

• Basics of Univariate, Bivariate & Multivariate plots
• Building plots with base r

Congratulations on completing this module!